An Effective Approach of Sentiment Analysis for Price Prediction
In today’s digital world, there has been an enormous amplification in the generation of user data. Due to progression in the internet-based applications, there is an exponential increase in social media users. Millions of users share their various opinions every day. Due to this fact, social media has become a powerful tool for communication and a rich source of opinion data. This opinion data is being used for prediction. This concept leads to the development of a new era of research called sentiment analysis. Sentiment Analysis, also known as opinion mining which uses new technologies and algorithms to collect and analyze opinions about various products and services. The main objective of this paper is to use sentiment analysis and machine learning in predicting stock prices. This functionality assists investors in predicting the stock market. The stock market may not be predicted accurately relaying only on the analysis of historical prices. To improve the prediction we added the sentiment scores along with prices by performing sentiment analysis based on opinions shared by different twitter users. Though predicting the Stock market is a tedious task and there exist different ways to predict it. We implemented a new approach of sentiment analysis by using the tensor flow platform and Twitter API which is used to get tweets. Our results shows that the proposed scheme exhibits better accuracy than the existing methods in predicting the stock market.
Keywords: Machine Learning, Prediction, Sentiment Analysis, Twitter API.